FIG. 1 illustrates a simplified block diagram of an LTE network 100. The LTE network 100 comprises a core network 102 and a radio access network (RAN) 104. The RAN 104 comprises a plurality of access nodes, which are commonly referred to a evolved NodeBs (eNodeBs) 108, providing service to user equipments (UEs) 110 in respective cells of the LTE network 100. The eNodeBs 108 communicate with the UEs 110 over the air interface and provide connection to the core network 102. Although the present invention is described in the context of an LTE network, those skilled in the art will appreciate that the invention can be used in other communication networks based on the communication standards, including without limitation Wideband Code Division Multiple Access (WCDMA) networks.
In an LTE system, PDCCHs are transmitted over radio resources that are shared between several UEs. The UE is specified as having to monitor four aggregation levels, specifically, 1, 2, 4, and 8, for a UE-specific search space and two aggregation levels, specifically, 4 and 8, for a common search space.
Third Generation Partnership Project (3GPP) Technical Specification (TS) 36.213 (Version 8), Section 9.1 explains the UE procedure for determining physical downlink control channel assignment. In particular, Section 9.1.1 (PDCCH assignment procedures) discusses a search space Sk(L) at an aggregation level Lε{1,2,4,8} which is defined by a contiguous set of Control Channel Elements (CCEs) given by(Zk(L)+i)mod NCCE,k  (1)where NCCE,k is the total number of CCEs in the control region of subframe k, Zk(L) defines the start of the search space, i=0, 1, . . . , M(L)·L−1 and M(L) is the number of PDCCHs to monitor in the given search space. Each CCE contains 36 Quadrature Phase Shift Keying (QPSK) modulation symbols. The value of M(L) is specified by Table 1 and disclosed in 3GPP TS 36.213, is shown below.
TABLE 1M(L) vs. Aggregation LevelNumber ofSearch space Sk(L)PDCCHAggregationSizecandidatesTypelevel L[in CCEs]M(L)UE-166specific21264828162Common41648162
With this definition, search space for different aggregation levels may overlap with each other regardless of system bandwidth. Specifically, UE-specific search space and common search space may overlap. In addition, the search spaces for different aggregation levels may overlap. For example, Table 2 below illustrates an example of such an overlap. Table 2 illustrates the example where NCCE,k=9, Zk(L)={1, 6, 4, 0} for L={1, 2, 4, 8}, respectively.
TABLE 2Example of Overlapping Aggregation LevelsSearch space Sk(L)AggregationTypeLevel LPDCCH candidatesUE-1{1}, {2}, {3}, {4}, {5}, {6}Specific2{6, 7}, {8, 0}, {1, 2}, {3, 4}, {5, 6},{7, 8}4{4, 5, 6, 7}, {8, 0, 1, 2}8{0, 1, 2, 3, 4, 5, 6, 7}, {8, 0, 1, 2, 3,4, 5, 6}Common4{0, 1, 2, 3}, {4, 5, 6, 7}, {8, 0, 1, 2},{3, 4, 5, 6}8{0, 1, 2, 3, 4, 5, 6, 7}, {8, 0, 1, 2, 3,4, 5, 6}
A PDCCH transmission employs circular buffer based rate matching for rate 1/3 Tail-biting Convolutional code. Due to repetition of coded bits and search space overlapping between different aggregation levels, multiple aggregation levels may pass the Cycle Redundancy Check (CRC) checking.
In addition, due to circular-buffer based rate matching, for a given aggregation size (2, 4 or 8), coded bits may start to repeat themselves after the 1st CCE. FIGS. 2A and 2B are simplified block diagrams illustrating CCE repetition examples in an existing telecommunications system. FIGS. 2A and 2B illustrate examples for a particular payload size (i.e., 48 bits). FIG. 2A illustrates a payload having a plurality of CCEs 200 having an aggregation size 4 with 2 repetitions. Each repetition starts at the same location in the circular buffer. FIG. 2B illustrates a payload with a plurality of CCEs 202 having an aggregation size 8. With an aggregation size of 8, there are four repetitions with each repetition starting at the same location in the circular buffer.
In general, the necessary condition to have confusing levels is shown in:N×k=24×m  (2)where N is the ambiguous payload size and m and k are both integers. Since the UE is not required to decode PDCCH with a code rate higher than 0.75, N should be no more than 54×(8−m). For example, when N=48, m=2 k, k may take a value of 1, 2, or 4. In such an example, any combination of {1,2,4,8} may create confusing (2 or more) aggregations levels. Since the LTE PDCCH payload contains information bits and the corresponding 16-bit CRC, the payload size is no less than 20 bits. An exhaustive list of all problematic sizes applicable to the LTE system is:{20,21,24,28,30,32,36,40,42,48,60,72,96,120}  (3)
Due to coded bits repetition and search space overlapping between different aggregation sizes, multiple aggregation sizes may pass the CRC checking. Since the 1st CCE of the PDCCH is linked to the uplink Acknowledgement/Negative Acknowledgement (ACK/NACK) resource for dynamic scheduling, the UE may send its ACK/NACK in a different resource, which is unknown by the Node-B (i.e., multiple ACK/NACK resources are possible). As such, there may be confusion in the Uplink (UL) ACK/NAK resource location mapped from the 1st CCE of the corresponding PDCCH grants, when two or more PDCCH decoding candidates from different aggregation levels have different lowest CCE indices. The potentially wrong UL ACK/NAK resource location not only creates unnecessary UL interference, it also impacts downlink throughput, especially for high geometry UEs.
There have been a large number of solutions to remedy these problems. In one solution, two bits are added in each PDCCH format to indicate the aggregation size. This simple solution would allow the UE to verify the correctness of the aggregation size. However, this solution increases the overhead on the PDCCH and reduces the coverage of these important system signals.
In another existing solution, for different aggregation sizes, a different CRC mask or scrambling codes is applied. This clearly increases UE decoding complexity. In addition, the additional scrambling operations for CRC (e.g., various UE identifications, Transmit antenna selection mask, and the proposal for aggregation level specific masks) lead to a higher CRC false detection probability. Thus, this solution does not address the problems associated with detection reliability.
In another existing solution, an evolved Node B (eNodeB) attempts to decode a UE's ACK/NACK at all possible locations. The eNodeB has no knowledge of whether a UE chooses the correct aggregation level for PDCCH transmission for ambiguous PDCCH payload sizes. The eNodeB may choose to detect UL ACK/NACK for a given UE on all possible aggregation levels. However, it not only creates additional implementation complexity, but, more importantly, it cannot guarantee correct detection. First, the eNodeB has to ensure no UL ACK/NAK collision is possible. This imposes a severe scheduling restriction as different UEs should not have overlapped search space. This is very difficult, if not impossible, to satisfy in reality due to system load. Moreover, the eNodeB cannot presume specific ACK/NAK statistics for certain aggregation levels due to channel conditions, a Hybrid Automatic Repeat Request (HARQ) termination target, and imperfect power control. Finally, multiple hypotheses inevitably provide a negatively impact UL ACK/NAK detection performance. Thus, it is not practical for utilizing an eNodeB to address the aforementioned problems.
To determine the actual aggregation level for a PDCCH transmission, a UE may use various approaches. The UE may use modulated symbols to determine the energy for all possible confusing combinations of CCEs. However, this approach is very unreliable because of interference from other cells. In another approach, the PDCCH may be re-encoded. The UE may decode bits to re-encode the PDCCH and determine the Signal-to-Noise Ratio (SNR) of all possible confusing combinations of CCEs. This approach is more reliable, but is very complicated. Alternately, the UE may perform a CRC check for each segment which contains an integer multiple of repetition of coded bits and an integer multiple of CCEs. There is no guarantee that each segment has the same CRC check result. Thus, complicated decision logic has to be devised. In addition, this approach inevitably increases the number of blind PDCCH decodes significantly. Thus, unless a very complicated implementation is adopted, a solution is not easily implemented. Alternatively, to bypass the above complicated implementations, the aggregation levels may be selected. For example, among all the aggregation levels with positive CRC checks, the highest or lowest aggregation level may be selected. In either case, the implementations are subject to non-negligible false alarm (of choosing the incorrect aggregation level) probability.
In another existing solution, zero padding may be applied to those PDCCH with the “troubled” payload sizes. Since, there are so many troubled payload sizes, this solution requires complicated receiver blind decoding logic.